The goal of this work is to create, validate, and apply an in silico model and tool to predict metabolites that are differentially accumulated in cancer. I is known that metabolites can broadly impact cellular behavior and growth outside of their roles as biosynthetic intermediates, and metabolism is being increasingly recognized as a potential target for cancer therapeutics. We believe that changes in concentration of some metabolites in cancer cells may have an active role in the progression of the disease rather than being just a side effect or consequence of other changes, such that the ability to predict these changes could result in the development of entirely new avenues of metabolism-focused cancer treatment. We have begun to develop an in silico model and tool, named CoMet, to make such predictions. In preliminary work using lymphoblasts, CoMet has successfully identified antiproliferative metabolites, though the accuracy of its predictions of metabolite levels, and its applicability to other types of cancer, is uncertain. To this end, the first aim of this proposal i to improve CoMet by integrating detailed biological data and using experimental validation results to refine its predictions. To perform our experimental validations, we will use a cutting-edge analytical technique (two-dimensional gas chromatography coupled to mass spectrometry, or GCxGC-MS) to measure the levels of metabolites in cancerous and normal cells and compare these results to predictions made by CoMet. Our second aim is to test the validity of CoMet's predictions of down-regulated and antiproliferative metabolites in multiple types of cancer, and to use these results to further refine CoMet's methodology. Our final aim is to measure the metabolic impact of using metabolites as antiproliferatives, since we suspect that they are having a substantial impact on cellular metabolism. This will allow us to generate hypotheses on their mechanisms of action. This work is a significant step towards gaining a predictive understanding of the metabolic differences between normal and cancerous cells, and of the regulatory roles metabolites play in cancer proliferation and progression. Predicting and understanding these changes would allow for the rational development of drugs that target cancer metabolism, and for advancement of the idea of metabolites that themselves serve as anticancer agents. By attacking such a fundamental aspect of cancer, this work could have a significant and broad long-term impact on cancer mortality and the quality of life of cancer patients. PUBLIC HEALTH RELEVANCE: This work aims to develop a model and computational tool to predict which metabolism intermediates are accumulated or depleted in cancerous cells. Predicting and understanding these changes would allow for the targeted development of drugs that attack cancer metabolism and thus limit cancer growth and progression. Additionally, these molecules may themselves play roles as natural anticancer agents.